Correction of Inertial Navigation System Drift Errors for an Autonomous Land Vehicle Using Optical Radar Terrain Data

摘要

It is generally agreed that some type of vision system capable of providing a preview of terrain is an important attribute of a driverless vehicle. One approach to providing this capability is to use active images such as those obtained by sonar or radar. This thesis is concerned with a computer simulation study of an approach to data processing for range images obtained from an optical radar system using a scanning laser beam. The system studied is modeled after the ERIM scanner mounted on the adaptive Suspension Vehicle walking machine developed at Ohio State University. Both the problem of register successive images in the presence of vehicle motion and of optimally averaging such images to obtain more accurate terrain elevation data are investigated in the thesis.